A methodology to determine the optimal quadrat size for desert vegetation surveying based on unmanned aerial vehicle (UAV) RGB photography

Quadrat sampling is one of the most widely accepted methods for conducting vegetation surveys over the world for centuries. However, it is difficult to determine an optimal size to adequately represent the community compositions in quadrat samplings. Traditional labour-intensive census-based quadrat sampling is also very time consuming and insufficient to represent spatial community characters outside of the quadrat extent. In order to improve the above deficiencies, an unmanned aerial vehicle (UAV)/red-green-blue (RGB) photography based vegetation survey methodology was proposed in this study. The essential steps in the proposed method include: 1) obtaining high spatial resolution optical images from the UAV; 2) extracting species based on the orthographic image after mosaic; 3) performing statistics on a given species within the image through different quadrat sizes; 4) determining an optimal quadrat size according to the changing trend of the statistical data. In addition, a case study applied this proposed methodology was conducted in a desert area in Northwest China, whereAmmopiptanthus mongolicusandZygophyllum xanthoxylondominated. The results show that the remote sensing UAV method could obtain RGB images and orthoimage with flexible control. The statistical data of species density decreased with the increase of quadrat size, but the values changed slightly after 20 m x 20 m, which was larger than the typical quadrat size (10 m x 10 m) used to investigate shrubs. An analysis based on the relationship among species density, plants distribution, and quadrat size further indicated the reasonability of determining the optimal size. Based on the results, it is concluded that a minimum quadrat size of 20 m x 20 m should be adopted for investigating the density and spatial pattern characteristics ofA. mongolicusandZ. xanthoxylon, and the proposed UAV-based method provides an alternative for vegetation survey with high efficiency and accuracy.